xarray dependency migration

CHANGELOG.md

@@ -19,6 +19,8 @@ but cannot always guarantee backwards compatibility. Changes that may **break co
 
 **Dependencies**
 
+- Moved `xarray` from core dependencies to optional dependencies. [#2974](https://github.com/unit8co/darts/pull/2974) by [Jakub Chłapek](https://github.com/jakubchlapek).
+
 ### For developers of the library:
 
 ## [0.39.0](https://github.com/unit8co/darts/tree/0.39.0) (2025-11-23)

darts/ad/utils.py

@@ -657,28 +657,31 @@ def _plot_series(series, ax_id, linewidth, label_name, **kwargs):
     label_name
         Name that will appear in the legend.
     """
-    data_array = series.data_array(copy=False)
-    for i, c in enumerate(data_array.component[:10]):
-        comp = data_array.sel(component=c)
-
-        if comp.sample.size > 1:
-            central_series = comp.mean(dim="sample")
-            low_series = comp.quantile(q=0.05, dim="sample")
-            high_series = comp.quantile(q=0.95, dim="sample")
+    for i, c in enumerate(series.components[:10]):
+        comp_series = series[c]
+
+        if series.is_stochastic:
+            central_series = comp_series.mean(axis=2)
+            low_series = comp_series.quantile(q=0.05)
+            high_series = comp_series.quantile(q=0.95)
         else:
-            central_series = comp
+            central_series = comp_series
 
         label_to_use = (
             (label_name + ("_" + str(i) if len(series.components) > 1 else ""))
             if label_name != ""
-            else "" + str(str(c.values))
+            else "" + str(c)
         )
 
         central_series.plot(ax=ax_id, linewidth=linewidth, label=label_to_use, **kwargs)
 
-        if comp.sample.size > 1:
+        if series.is_stochastic:
             ax_id.fill_between(
-                series.time_index, low_series, high_series, alpha=0.25, **kwargs
+                series.time_index,
+                low_series.values().flatten(),
+                high_series.values().flatten(),
+                alpha=0.25,
+                **kwargs,
             )
 
 

darts/dataprocessing/dtw/_plot.py

@@ -1,7 +1,6 @@
 from typing import Union
 
 import numpy as np
-import xarray as xr
 from matplotlib import pyplot as plt
 
 
@@ -142,23 +141,17 @@ def plot_alignment(
     series1 += series1_y_offset
     series2 += series2_y_offset
 
-    xa1 = series1.data_array(copy=False)
-    xa2 = series2.data_array(copy=False)
-
     path = self.path()
     n = len(path)
 
-    time_dim1 = series1.time_dim
-    time_dim2 = series2.time_dim
-
-    x_coords1 = np.array(xa1[time_dim1], dtype=xa1[time_dim1].dtype)[path[:, 0]]
-    x_coords2 = np.array(xa2[time_dim2], dtype=xa2[time_dim2].dtype)[path[:, 1]]
+    x_coords1 = series1.time_index[path[:, 0]]
+    x_coords2 = series2.time_index[path[:, 1]]
 
     y_coords1 = series1.univariate_values()[path[:, 0]]
     y_coords2 = series2.univariate_values()[path[:, 1]]
 
     if series1.has_datetime_index:
-        x_dtype = xa1[time_dim1].dtype
+        x_dtype = series1.time_index.dtype
         x_nan = np.datetime64("NaT")
     else:
         x_dtype = np.float64
@@ -175,8 +168,7 @@ def plot_alignment(
     y_coords[1::3] = y_coords2
     y_coords[2::3] = np.nan
 
-    arr = xr.DataArray(y_coords, dims=["value"], coords={"value": x_coords})
-    xr.plot.line(arr, x="value", **args_line)
-
     series1.plot(**args_series1)
     series2.plot(**args_series2)
+
+    plt.plot(x_coords, y_coords, **args_line)

darts/tests/dataprocessing/dtw/test_dtw.py

@@ -154,13 +154,9 @@ def test_itakura_window(self):
             assert 1 > dist
 
     def test_warp(self):
-        # Support different time dimension names
-        xa1 = self.series1.data_array().rename({"time": "time1"})
-        xa2 = self.series2.data_array().rename({"time": "time2"})
-
         static_covs = pd.DataFrame([[0.0, 1.0]], columns=["st1", "st2"])
-        series1 = TimeSeries.from_xarray(xa1).with_static_covariates(static_covs)
-        series2 = TimeSeries.from_xarray(xa2).with_static_covariates(static_covs)
+        series1 = self.series1.with_static_covariates(static_covs)
+        series2 = self.series2.with_static_covariates(static_covs)
         series1_copy = series1.copy()
         series2_copy = series2.copy()
 

darts/tests/test_timeseries.py

@@ -7,7 +7,6 @@
 import numpy as np
 import pandas as pd
 import pytest
-import xarray as xr
 from scipy.stats import kurtosis, skew
 
 from darts import TimeSeries, concatenate, slice_intersect
@@ -17,7 +16,7 @@
     quantile_names,
 )
 from darts.utils.timeseries_generation import constant_timeseries, linear_timeseries
-from darts.utils.utils import expand_arr, freqs, generate_index
+from darts.utils.utils import XARRAY_AVAILABLE, expand_arr, freqs, generate_index
 
 TEST_BACKENDS = ["pandas"]
 
@@ -65,14 +64,17 @@ def test_creation(self):
         )
         assert ts.components.tolist() == ["a", "b", "a_1", "a_2", "b_1"]
 
-        # creation using from_xarray()
+    @pytest.mark.skipif(not XARRAY_AVAILABLE, reason="xarray required")
+    def test_xarray_creation(self):
+        import xarray as xr
+
         ar = xr.DataArray(
             np.random.randn(10, 2, 1),
             dims=("time", "component", "sample"),
             coords={"time": self.times, "component": ["a", "b"]},
-            name="time series",
         )
-        _ = TimeSeries.from_xarray(ar)
+        ts = TimeSeries.from_xarray(ar)
+        assert ts.components.tolist() == ["a", "b"]
 
     def test_from_times_and_values(self):
         # Test creation from times and values
@@ -401,7 +403,10 @@ def test_univariate_component(self):
             # only the right static covariate column should be retained
             assert univ_series.static_covariates.sum().sum() == 1.1
 
+    @pytest.mark.skipif(not XARRAY_AVAILABLE, reason="requires xarray")
     def test_column_names(self):
+        import xarray as xr
+
         # test the column names resolution
         columns_before = [
             ["0", "1", "2"],
@@ -810,14 +815,17 @@ def test_ops(self):
             # Cannot divide by 0.
             self.series1 / 0
 
-    def test_ops_array(self):
+    @pytest.mark.skipif(not XARRAY_AVAILABLE, reason="xarray required")
+    def test_ops_xarray(self):
         # can work with xarray directly
         series2_x = self.series2.data_array(copy=False)
         assert self.series1 + self.series2 == self.series1 + series2_x
         assert self.series1 - self.series2 == self.series1 - series2_x
         assert self.series1 * self.series2 == self.series1 * series2_x
         assert self.series1 / self.series2 == self.series1 / series2_x
         assert self.series1**self.series2 == self.series1**series2_x
+
+    def test_ops_ndarray(self):
         # can work with ndarray directly
         series2_nd = self.series2.all_values(copy=False)
         assert self.series1 + self.series2 == self.series1 + series2_nd
@@ -1277,6 +1285,7 @@ def test_fillna_value(self):
             assert not np.isnan(series_no_nan.all_values(copy=False)).any()
             assert series_1 == series_no_nan
 
+    @pytest.mark.skipif(not XARRAY_AVAILABLE, reason="xarray required")
     def test_resample_timeseries(self):
         # 01/01/2013 -> 10/01/2013, one value per day: 0 1 2 3 ... 9
         times = pd.date_range("20130101", "20130110")
@@ -2404,7 +2413,7 @@ def helper_test_prepend(test_series: TimeSeries):
 def helper_test_prepend_values(test_series: TimeSeries):
     # reconstruct series
     seriesA, seriesB = test_series.split_after(pd.Timestamp("20130106"))
-    arrayA = seriesA.data_array().values
+    arrayA = seriesA._values
     prepended = seriesB.prepend_values(arrayA)
     assert prepended == test_series
     assert test_series.time_index.equals(prepended.time_index)

darts/tests/test_timeseries_static_covariates.py

@@ -17,7 +17,7 @@
     STATIC_COV_TAG,
 )
 from darts.utils.timeseries_generation import linear_timeseries
-from darts.utils.utils import generate_index
+from darts.utils.utils import XARRAY_AVAILABLE, generate_index
 
 TEST_BACKENDS = ["pandas"]
 
@@ -92,7 +92,8 @@ def test_ts_from_x(self, tag, tmpdir_module):
         ts, x = setup_tag(tag, ts)
         kwargs = {tag: x}
 
-        self.helper_test_transfer(tag, ts, TimeSeries.from_xarray(ts.data_array()))
+        if XARRAY_AVAILABLE:
+            self.helper_test_transfer(tag, ts, TimeSeries.from_xarray(ts.data_array()))
         self.helper_test_transfer(
             tag, ts, TimeSeries.from_dataframe(ts.to_dataframe(), **kwargs)
         )
@@ -922,7 +923,8 @@ def test_ts_methods(self, tag):
         self.helper_test_transfer(tag, ts, ts.diff())
         self.helper_test_transfer(tag, ts, ts.univariate_component(0))
         self.helper_test_transfer(tag, ts, ts.map(lambda x: x + 1))
-        self.helper_test_transfer(tag, ts, ts.resample(ts.freq))
+        if XARRAY_AVAILABLE:
+            self.helper_test_transfer(tag, ts, ts.resample(ts.freq))
         self.helper_test_transfer(tag, ts, ts[:5].append(ts[5:]))
         self.helper_test_transfer(tag, ts, ts.append_values(ts.all_values()))